nep-for New Economics Papers
on Forecasting
Issue of 2013‒10‒18
fourteen papers chosen by
Rob J Hyndman
Monash University

  1. Uncertainty and heterogeneity in factor models forecasting By Matteo Luciani; Libero Monteforte
  2. Currency forecast errors at times of low interest rates: evidence from survey data on the Yen/Dollar exchange rate By Ronald MacDonald; Jun Nagayasu
  3. Forecasting aggregate demand: analytical comparison of top-down and bottom-up approaches in a multivariate exponential smoothing framework By Giacomo Sbrana; Andrea Silvestrini
  4. Short-term forecasting with business surveys: Evidence for German IHK data at federal state level By Wenzel, Lars; Wolf, André
  5. Inflation Targeting and Macroeconomic Stability with Heterogeneous Inflation Expectations By Gilberto Tadeu Lima; Mark Setterfield, Jaylson Jair da Silveira
  6. Forecasting the potential of Danish biogas production: spatial representation of Markov chains By Mikkel Bojesen; Hans Skov-Petersen; Morten Gylling
  7. Estimation Errors in Input-Output Tables and Prediction Errors in Computable General Equilibrium Analysis By Nobuhiro Hosoe
  8. Nowcasting Czech GDP in Real Time By Marek Rusnak
  9. Detecting and Forecasting Large Deviations and Bubbles in a Near-Explosive Random Coefficient Model By Banerjee, Anurag N.; Chevillon, Guillaume; Kratz, Marie
  10. Some further estimations for: Voting and economic factors in French elections for the European Parliament By Antoine Auberger
  11. Nonlife Ratemaking and Risk Management with Bayesian Additive Models for Location, Scale and Shape By Nadja Klein; Michel Denuit; Stefan Lang; Thomas Kneib
  12. Identifying long-run risks: a bayesian mixed-frequency approach By Frank Schorfheide; Dongho Song; Amir Yaron
  13. A Dynamic Duverger's Law By Jean Guillaume Forand; Vikram Maheshri
  14. The lighting fixtures market in Asia Pacific By Aurelio Volpe

  1. By: Matteo Luciani (Université libre de Bruxelles); Libero Monteforte (Bank of Italy)
    Abstract: In this paper, we exploit the heterogeneity in the forecasts obtained by estimating different factor models to measure forecast uncertainty. Our approach is simple and intuitive. It consists first in selecting all the models that outperform some benchmark model, and then in constructing an empirical distribution of the forecasts produced by them. We interpret this distribution as a measure of uncertainty. We illustrate our methodology by means of a forecasting exercise using a large database of Italian data from 1982 to 2009.
    Keywords: factor models, model uncertainty, forecast combination, density forecast
    JEL: C13 C32 C33 C52 C53
    Date: 2013–09
  2. By: Ronald MacDonald (Adam Smith Business School, University of Glasgow); Jun Nagayasu (Faculty of Engineering, Information and Systems, University of Tsukuba)
    Abstract: Using survey expectations data and Markov-switching models, this paper evaluates the characteristics and evolution of investors’ forecast errors about the yen/dollar exchange rate. Since our model is derived from the uncovered interest rate parity (UIRP) condition and our data cover a period of low interest rates, this study is also related to the forward premium puzzle and the currency carry trade strategy. We obtain the following results. First, with the same forecast horizon, exchange rate forecasts are homogeneous among different industry types, but within the same industry, exchange rate forecasts differ if the forecast time horizon is different. In particular, investors tend to undervalue the future exchange rate for long term forecast horizons; however, in the short run they tend to overvalue the future exchange rate. Second, while forecast errors are found to be partly driven by interest rate spreads, evidence against the UIRP is provided regardless of the forecasting time horizon; the forward premium puzzle becomes more significant in shorter term forecasting errors. Consistent with this finding, our coefficients on interest rate spreads provide indirect evidence of the yen carry trade over only a short term forecast horizon. Furthermore, the carry trade seems to be active when there is a clear indication that the interest rate will be low in the future.
    Keywords: Currency forecast errors, uncovered interest parity, forward premium puzzle, carry trade, Markov-switching model
    JEL: F3
    Date: 2013–10
  3. By: Giacomo Sbrana (Rouen Business School); Andrea Silvestrini (Bank of Italy)
    Abstract: Forecasting aggregate demand is a crucial matter in all industrial sectors. In this paper, we provide the analytical prediction properties of top-down (TD) and bottom-up (BU) approaches when forecasting aggregate demand, using multivariate exponential smoothing as demand planning framework. We extend and generalize the results obtained by Widiarta, Viswanathan and Piplani (2009) by employing an unrestricted multivariate framework allowing for interdependency between the variables. Moreover, we establish the necessary and sufficient condition for the equality of mean squared errors (MSEs) of the two approaches. We show that the condition for the equality of MSEs also holds even when the moving average parameters of the individual components are not identical. In addition, we show that the relative forecasting accuracy of TD and BU depends on the parametric structure of the underlying framework. Simulation results confirm our theoretical findings. Indeed, the ranking of TD and BU forecasts is led by the parametric structure of the underlying data generation process, regardless of possible misspecification issues.
    Keywords: top-down and bottom-up forecasting, multivariate exponential smoothing.
    JEL: C32 C43
    Date: 2013–09
  4. By: Wenzel, Lars; Wolf, André
    Abstract: We investigate the performance of the IHK business climate indices as forecasting tools within a growth framework at the level of four federal states in Northern Germany. In doing this, we match quarterly index scores with estimates of quarterly production data, generated through a Chow-Lin procedure. Estimating the model reveals strong linkages of the index scores to short-term output growth at the regional level, even after controlling for prior information on the position in the business cycle as well as for nation-wide fluctuations. Examining the forecasting accuracy of our model by means of out-of-sample predictions confirms these results: the model clearly outperforms an autoregressive benchmark. This can to a large part be traced back to information conveyed by the IHK index. --
    Date: 2013
  5. By: Gilberto Tadeu Lima; Mark Setterfield, Jaylson Jair da Silveira
    Abstract: Drawing on an extensive empirical literature that suggests persistent and time-varying heterogeneity in inflation expectations, this paper embeds two inflation forecasting heuristics – one based on the current rate of inflation, the second anchored to the official inflation target – in a simple macrodynamic model. Decision makers switch between these forecasting heuristics based on satisficing evolutionary dynamics. We show that convergence towards an equilibrium consistent with the level of output and rate of inflation targeted by policy makers is achieved regardless of whether or not the satisficing evolutionary dynamics that guide the choices agents make between inflation forecasting strategies are subject to noise. We also show that full credulity – a situation where all agents eventually use the forecasting heuristic based on the target rate of inflation – is neither a necessary condition for realization of the inflation target, nor an inevitable consequence of the economy’s achievement of this target. These results demonstrate that uncertainty in decision making resulting in norm-based inflation expectations that are both heterogeneous and time-varying need not thwart the successful conduct of macroeconomic policy.
    Keywords: Inflation targeting; macroeconomic stability; heterogeneous expected inflation; satisficing evolutionary dynamics
    JEL: C73 E12 E52
    Date: 2013–10–01
  6. By: Mikkel Bojesen (Department of Food and Resource Economics, University of Copenhagen); Hans Skov-Petersen (Department of Geosciences and Natural Resource Management, University of Copenhagen); Morten Gylling (Department of Food and Resource Economics, University of Copenhagen)
    Abstract: This paper forecasts the spatial distribution of Danish husbandry production from 2009 until 2025. The study builds on a time series data set (1999 – 2009) for the number of livestock units (including piglets, finishers, sows, dairy cattle and young stock) measured by 1km2 grid cells for the whole of Denmark. A Markov Chain Model (MCM) was applied to estimate transition probabilities for the future livestock intensity composition, divided into state classes. Neighbouring effects between grid cells were not included. The modelled transition probabilities fit the data very well in all state classes, except for those farms which are in the largest state class. Regional differences in development trends were documented. The strategic objective of the model is to provide data for the spatial assessment of the potential of biogas production which can form the basis for a location analysis for future biogas plants.
    Keywords: structural changes, Markov Chain Models, biogas, agriculture
    JEL: Q47 Q16 Q21
    Date: 2013–10
  7. By: Nobuhiro Hosoe (National Graduate Institute for Policy Studies)
    Abstract: We used 1995-2000-2005 linked input-output (IO) tables for Japan to examine estimation errors of updated IO tables and the resulting prediction errors in computable general equilibrium (CGE) analysis developed with updated IO tables. As we usually have no true IO tables for the target year and therefore need to estimate them, we cannot evaluate estimation errors of updated IO tables without comparing the updated ones with true ones. However, using the linked IO tables covering three different years enables us to make this comparison. Our experiments showed that IO tables estimated with more detailed and recent data contained smaller estimation errors and led to smaller quantitative prediction errors in CGE analysis. Despite the quantitative prediction errors, prediction was found to be qualitatively correct. As for the performance of updating techniques of IO tables, a cross-entropy method often outperformed a least-squares method in IO estimation with only aggregate data for the target year but did not necessarily outperform the least-squares method in CGE prediction.
    Date: 2013–10
  8. By: Marek Rusnak
    Abstract: The prominent measure of the current state of the Czech economy, gross domestic product (GDP), is available only with a significant lag of roughly 70 days. In this paper, we employ a Dynamic Factor Model (DFM) to nowcast Czech GDP in real time. Using multiple vintages of historical data and taking into account the publication lags of various monthly indicators, we evaluate the real-time performance of the DFM over the 2005–2012 period. The results suggest that the accuracy of model-based nowcasts is comparable to that of the judgmental nowcasts of the Czech National Bank (CNB). Our results also suggest that foreign variables are crucial for the accuracy of the model, while omitting financial and confidence indicators does not worsen the nowcasting performance. Finally, we show how releases of new data can be viewed through the lens of the dynamic factor model.
    Keywords: Dynamic factor model, GDP, nowcasting, real-time data.
    JEL: C53 C82 E52
    Date: 2013–07
  9. By: Banerjee, Anurag N. (Durham University Business School); Chevillon, Guillaume (ESSEC Business School); Kratz, Marie (ESSEC Business School et Mathématiques appliquées Paris 5 (MAP5))
    Abstract: This paper proposes a Near Explosive Random-Coefficient autoregressive model for asset pricing which accommodates both the fundamental asset value and the recurrent presence of autonomous deviations or bubbles. Such a process can be stationary with or without fat tails, unit-root nonstationary or exhibit temporary exponential growth. We develop the asymptotic theory to analyze ordinary least-squares (OLS) estimation. One important theoretical observation is that the estimator distribution in the random coefficient model is qualitatively different from its distribution in the equivalent fixed coefficient model. We conduct recursive and full-sample inference by inverting the asymptotic distribution of the OLS test statistic, a common procedure in the presence of localizing parameters. This methodology allows to detect the presence of bubbles and establish probability statements on their apparition and devolution. We apply our methods to the study of the dynamics of the Case-Shiller index of U.S. house prices. Focusing in particular on the change in the price level, we provide an early detection device for turning points of booms and bust of the housing market.
    Keywords: Bubbles; Random Coefficient Autoregressive Model; Local Asymptotics; Asset Prices
    JEL: C22 C53 C58 G12
    Date: 2013–09
  10. By: Antoine Auberger (IRGEI - Institut de Recherche sur la Gouvernance et l'Economie des Institutions - Université Paris II - Panthéon-Assas)
    Abstract: The purpose of this note is to complete the estimates made in Auberger (2012) for French presidential elections. We study the influence of the local unemployment on the vote for French presidential elections. We build another variable taking into account the responsibility of the incumbent President for the economic situation after a cohabitation period. We also make estimates for the second-round vote of French presidential elections (without the 2002 French presidential election or with an estimated vote for this election). We show that over the 1988-2007 period (without 2002), it is not necessary to take into account the influence of cohabitation periods on the responsibility of the government in relation to the economic situation.
    Keywords: Vote functions ; French elections ; European Parliament ; Election forecasting ; Local unemployment
    Date: 2013–09–12
  11. By: Nadja Klein; Michel Denuit; Stefan Lang; Thomas Kneib
    Abstract: Generalized additive models for location, scale and shape define a flexible, semi-parametric class of regression models for analyzing insurance data in which the exponential family assumption for the response is relaxed. This approach allows the actuary to include risk factors not only in the mean but also in other parameters governing the claiming behavior, like the degree of residual heterogeneity or the no-claim probability. In this broader setting, the Negative Binomial regression with cell-specific heterogeneity and the zero-inflated Poisson regression with cell-specific additional probability mass at zero are applied to model claim frequencies. Models for claim severities that can be applied either per claim or aggregated per year are also presented. Bayesian inference is based on efficient Markov chain Monte Carlo simulation techniques and allows for the simultaneous estimation of possible nonlinear effects, spatial variations and interactions between risk factors within the data set. To illustrate the relevance of this approach, a detailed case study is proposed based on the Belgian motor insurance portfolio studied in Denuit and Lang (2004).
    Keywords: overdispersed count data, mixed Poisson regression, zero-inflated Poisson, Negative Binomial, zero-adjusted models, MCMC, probabilistic forecasts
    Date: 2013–10
  12. By: Frank Schorfheide; Dongho Song; Amir Yaron
    Abstract: We develop a nonlinear state-space model that captures the joint dynamics of consumption, dividend growth, and asset returns. Building on Bansal and Yaron (2004), our model consists of an economy containing a common predictable component for consumption and dividend growth and multiple stochastic volatility processes. The estimation is based on annual consumption data from 1929 to 1959, monthly consumption data after 1959, and monthly asset return data throughout. We maximize the span of the sample to recover the predictable component and use high-frequency data, whenever available, to efficiently identify the volatility processes. Our Bayesian estimation provides strong evidence for a small predictable component in consumption growth (even if asset return data are omitted from the estimation). Three independent volatility processes capture different frequency dynamics; our measurement error specification implies that consumption is measured much more precisely at an annual than monthly frequency; and the estimated model is able to capture key asset-pricing facts of the data.
    Keywords: Nonlinear theories ; Dividends ; Consumption (Economics) ; Bayesian statistical decision theory
    Date: 2013
  13. By: Jean Guillaume Forand (Department of Economics, University of Waterloo); Vikram Maheshri (Department of Economics, University of Houston)
    Abstract: Electoral systems promote strategic voting and aect party systems. Duverger (1951) proposed that plurality rule leads to bi-partyism and proportional representation leads to multi-partyism. We show that in a dynamic setting, these static eects also lead to a higher option value for existing minor parties under plurality rule, so their incentive to exit the party system is mitigated by their future benets from continued participation. The predictions of our model are consistent with multiple cross-sectional predictions on the comparative number of parties under plurality rule and proportional representation. In particular, there could be more parties under plurality rule than under proportional representation at any point in time. However, our model makes a unique time-series prediction: the number of parties under plurality rule should be less variable than under proportional representation. We provide extensive empirical evidence in support of these results.
    JEL: C73 D72
    Date: 2013–10
  14. By: Aurelio Volpe (CSIL Centre for Industrial Studies)
    Abstract: The report analyzes the lighting fixtures market in Asia Pacific. It provides historical statistical data of Production, International trade and Market size 2007-2012 of the lighting fixtures industry for the 11 countries considered. Macroeconomic forecasts up to 2015 are also provided. Demand is broken down by product and by light source. An overview of the competitive system is outlined, with short company profiles, data on sales and market shares for the top lighting fixtures companies for each Asia Pacific country considered. International trade of of lighting fixtures and distribution channels are included. Addresses of about 460 lighting fixtures manufacturers and other key contacts are also included. Countries covered: Australia, Indonesia, Japan, Malaysia, New Zealand, Philippines, Singapore, South Korea, Taiwan, Thailand, Vietnam.
    JEL: L11 L22 L68
    Date: 2013–09

This nep-for issue is ©2013 by Rob J Hyndman. It is provided as is without any express or implied warranty. It may be freely redistributed in whole or in part for any purpose. If distributed in part, please include this notice.
General information on the NEP project can be found at For comments please write to the director of NEP, Marco Novarese at <>. Put “NEP” in the subject, otherwise your mail may be rejected.
NEP’s infrastructure is sponsored by the School of Economics and Finance of Massey University in New Zealand.